[Tockit-general] New Docco test version

[Peter Becker]

Hi all,

if some of you can spare some time: here is a new version of Docco you
could test for me:

http://kvo.itee.uq.edu.au/~pbecker/Docco-CVSbuild-20040406-0003.zip
(~1.5MB)

New features/fixes:
- File->Exit now closes properly (0.2 did not safe the index metainfo
properly)
- the "View" menu is now called "Diagram" and contains extra entries for
graphic export and printing
- nodes can be moved, standard left-drag uses the n-dim manipulator,
shift-left-drag moves single nodes
- the query field now has a drop-down list with the last queries (up to
ten). This is restored between sessions, as well as the last query.
- all session management went into the Java preferences API, which means
Registry on Windows machine, .java on UNIX/Linux. Advantages:
- user-specific
- no new settings after upgrades
- no write access to installation directory needed

I also updated the PDFbox and POI plugins:


http://kvo.itee.uq.edu.au/~pbecker/pdfboxPlugin-CVSbuild-20040406-0003.zip
(~2MB)

http://kvo.itee.uq.edu.au/~pbecker/poiPlugin-CVSbuild-20040406-0003.zip
(~1MB)

PDFBox is just a library-upgrade, they claim lots of bugfixes. It is
quite a bit larger now -- thanks to some new Unicode mappings it seems.
No idea if it is worth it -- experiments comparing the two versions are
welcome.

The old POI plugin started to fail on Word documents, it seem due to JRE
changes. Should work again, it also uses the Textmining.org text
extractor, which should get more text out of documents (tables etc.) and
shoudl work on Word 6.0 documents.

I'll try to give it a documentation update soon and then release it as a
new version unless someone founds bugs.

Regards,
Peter

[Bastian Wormuth]

Hi Peter,

here is my first bug report:

- I miss the option to recreate the diagram by hitting the "Submit"
button again, this was a nice option to change the diagrams layout.
I guess not intended, but nice anyway. It is still possible by
hitting "Enter", but not via the button, it remains grey. Nothing
serious.
- The PDF plugin still has the problem I mentioned earlier. I get just
one hit out of 52 on the BIS example, the "Schulz.pdf". Also "ext:pdf"
finds just that one. I tried it on another folder, same problem,
just 5 out of 15.

- A question: what about PDF/PS export ? Could I use the same plugins
as in TJ ?

Besides this - very nice, I like it very much, especially node
dragging ;-)

Cheers,

Bastian
--
Bastian Wormuth
Lauteschlägerstrasse 42
D - 64289 Darmstadt

Tel: ++49 (0) 6151/718039
Email: ***@mathematik.tu-darmstadt.de

Monday, April 05, 2004, 4:14:27 PM, you wrote:

PB> Hi all,

PB> if some of you can spare some time: here is a new version of Docco you
PB> could test for me:

PB> http://kvo.itee.uq.edu.au/~pbecker/Docco-CVSbuild-20040406-0003.zip
PB> (~1.5MB)

PB> New features/fixes:
- File->>Exit now closes properly (0.2 did not safe the index metainfo
PB> properly)
PB> - the "View" menu is now called "Diagram" and contains extra entries for
PB> graphic export and printing
PB> - nodes can be moved, standard left-drag uses the n-dim manipulator,
PB> shift-left-drag moves single nodes
PB> - the query field now has a drop-down list with the last queries (up to
PB> ten). This is restored between sessions, as well as the last query.
PB> - all session management went into the Java preferences API, which means
PB> Registry on Windows machine, .java on UNIX/Linux. Advantages:
PB> - user-specific
PB> - no new settings after upgrades
PB> - no write access to installation directory needed

PB> I also updated the PDFbox and POI plugins:


PB> http://kvo.itee.uq.edu.au/~pbecker/pdfboxPlugin-CVSbuild-20040406-0003.zip
PB> (~2MB)

PB> http://kvo.itee.uq.edu.au/~pbecker/poiPlugin-CVSbuild-20040406-0003.zip
PB> (~1MB)

PB> PDFBox is just a library-upgrade, they claim lots of bugfixes. It is
PB> quite a bit larger now -- thanks to some new Unicode mappings it seems.
PB> No idea if it is worth it -- experiments comparing the two versions are
PB> welcome.

PB> The old POI plugin started to fail on Word documents, it seem due to JRE
PB> changes. Should work again, it also uses the Textmining.org text
PB> extractor, which should get more text out of documents (tables etc.) and
PB> shoudl work on Word 6.0 documents.

PB> I'll try to give it a documentation update soon and then release it as a
PB> new version unless someone founds bugs.

PB> Regards,
PB> Peter


PB> -------------------------------------------------------
PB> This SF.Net email is sponsored by: IBM Linux Tutorials
PB> Free Linux tutorial presented by Daniel Robbins, President and CEO of
PB> GenToo technologies. Learn everything from fundamentals to system
PB> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
PB> _______________________________________________
PB> Tockit-general mailing list
PB> Tockit-***@lists.sourceforge.net
PB> https://lists.sourceforge.net/lists/listinfo/tockit-general

[Peter Becker]

Post by Bastian Wormuth
Hi Peter,
- I miss the option to recreate the diagram by hitting the "Submit"
button again, this was a nice option to change the diagrams layout.
I guess not intended, but nice anyway. It is still possible by
hitting "Enter", but not via the button, it remains grey. Nothing
serious.

It actually takes a little bit of code to disable the button. I can
remove that.

Post by Bastian Wormuth
- The PDF plugin still has the problem I mentioned earlier. I get just
one hit out of 52 on the BIS example, the "Schulz.pdf". Also "ext:pdf"
finds just that one. I tried it on another folder, same problem,
just 5 out of 15.

First test I did was running against the ICFCA04 CD. I got 3/4 hits, the
ToscanaJ manual failed. There seems to be some problem with the OOo
PDFs, even though the Elba manual worked. I attached some stuff to this
bug in PDFbox:
-
http://sourceforge.net/tracker/index.php?func=detail&aid=860027&group_id=78314&atid=552832

I could reproduce the problem with the BIS example, though.
Unfortunately it is a feature, not a bug -- see the last entry for 0.6.3
on this list:
- http://www.csh.rit.edu/~ben/projects/pdfbox/news.html

Acrobat Reader disallows copying text out of these documents, too --
although it does allow text selection, whatever that is good for if you
can't copy it.

So this is actually correct behaviour. Not that I really want it, though
:-( But I don't think I want to hack PDFbox either -- which shouldn't be
too hard (a diff between the source of 0.6.2 and 0.6.3 would most likely
help a lot), but while I disagree with this type of secretiveness, I
still think we should respect that.

Post by Bastian Wormuth
- A question: what about PDF/PS export ? Could I use the same plugins
as in TJ ?

Yes, the diagram export plugins are quite low-level: they operate on the
canvas package, which means any program using the canvas and the plugin
loader does automatically use them.

Post by Bastian Wormuth
Besides this - very nice, I like it very much, especially node
dragging ;-)

Thanks. What do you think about restoring the last query? I am still not
sure if I really want that, I just couldn't be bothered yesterday to
find a way to start with an empty text field. JComboBox is yet another
really bad part of Swing :-( Quite a bit of swearing involved in getting
there. :-)

Peter

[Peter Becker]

Two additions...

Post by Peter Becker

Post by Bastian Wormuth
Hi Peter,
- I miss the option to recreate the diagram by hitting the "Submit"
button again, this was a nice option to change the diagrams layout.
I guess not intended, but nice anyway. It is still possible by
hitting "Enter", but not via the button, it remains grey. Nothing
serious.

It actually takes a little bit of code to disable the button. I can
remove that.

Changed.

[..PDFbox, export plugins...]

Post by Peter Becker

Post by Bastian Wormuth
Besides this - very nice, I like it very much, especially node
dragging ;-)

Thanks. What do you think about restoring the last query? I am still
not sure if I really want that, I just couldn't be bothered yesterday
to find a way to start with an empty text field. JComboBox is yet
another really bad part of Swing :-( Quite a bit of swearing involved
in getting there. :-)

That one was actually easy in the morning (not that it changes my
opinion about JComboBox much). I can do it whatever way we like.

Peter

[Bastian Wormuth]

Tuesday, April 06, 2004, 2:04:26 AM, you wrote:

PB> Bastian Wormuth wrote:

PB> Thanks. What do you think about restoring the last query? I am still not
PB> sure if I really want that, I just couldn't be bothered yesterday to
PB> find a way to start with an empty text field. JComboBox is yet another
PB> really bad part of Swing :-( Quite a bit of swearing involved in getting
PB> there. :-)

No, it good, that would have been my next request, if you hadn't done
it yourself !

Regards,

Bastian

[Peter Becker]

PB> First test I did was running against the ICFCA04 CD. I got 3/4 hits, the
PB> ToscanaJ manual failed. There seems to be some problem with the OOo
PB> PDFs, even though the Elba manual worked. I attached some stuff to this
PB> -
PB> http://sourceforge.net/tracker/index.php?func=detail&aid=860027&group_id=78314&atid=552832
PB> I could reproduce the problem with the BIS example, though.
PB> Unfortunately it is a feature, not a bug -- see the last entry for 0.6.3
PB> - http://www.csh.rit.edu/~ben/projects/pdfbox/news.html
PB> Acrobat Reader disallows copying text out of these documents, too --
PB> although it does allow text selection, whatever that is good for if you
PB> can't copy it.
PB> So this is actually correct behaviour. Not that I really want it, though
PB> :-( But I don't think I want to hack PDFbox either -- which shouldn't be
PB> too hard (a diff between the source of 0.6.2 and 0.6.3 would most likely
PB> help a lot), but while I disagree with this type of secretiveness, I
PB> still think we should respect that.
... mmmh, so what to do ? PDFbox disallows copying text, but that's
not what we want, we just want to search for some text. Isn't there
any way around it ? Or couldn't we just use the old version ?

These particular PDFs are encoded in a way that text is not supposed to
be extracted -- neither by copy&paste nor by tools. It seems the old
PDFbox ignored that, the new one respects this. Of course there is most
likely a way around it, but I don't think it is important at the moment
-- and I won't remove the old plugin, which you can still use.

One thing I should do is adding better feedback, though. At the moment
you have to hack the run-docco.bat and turn the javaw call into java,
then check the output on the command line. Not nice, but I don't think
I'll fix that in this version.

Peter